BV ERCOT Gas Study Report March 2012

7/31/2019 BV ERCOT Gas Study Report March 2012

1/107

GASCURTAILMENTRISKSTUDY

Preparedfor

TheElectricReliabilityCouncilofTexas

MARCH2012

Black&VeatchHoldingCompany2011.Allrightsreserved.


2/107

TheElectricReliabilityCouncilofTexas|GASCURTAILMENTRISKSTUDY

BLACK&VEATCH|TableofContents i

TableofContentsTableofContents..........................................................................................................i

1.0ExecutiveSummary............................................................................................3

2.0Introduction..........................................................................................................8

2.1 Organization of this Report ...................................................................................... 9

3.0ReviewofHistoricalCurtailments..............................................................11

3.1 Data Availability & Sources .................................................................................... 12

3.2 Summary of Historical Curtailments & Causes .............................................. 13

3.3 Best Practices & Lessons to be Learned ............................................................ 16

4.0NaturalGasInfrastructure&Market.........................................................18

4.1 Interstate Pipelines................................................................................................... 18

4.2 Intrastate Pipelines................................................................................................... 19

4.3 Natural Gas Storage Facilities .............................................................................. 20

4.4 Role of Gas Compressors ........................................................................................ 21

4.5 Survey Results of Natural Gas Infrastructure Serving ERCOT

Generators ............................................................................................................................ 22

5.0 RiskAssessment -Approach&Assumptions........................................26

5.1 Summary of Approach ............................................................................................. 26

5.2 Tools and Software .................................................................................................... 29

5.3 Global Assumptions ................................................................................................... 31

6.0 RiskAssessment- Results.............................................................................35

6.1 Implications from Freezing Weather ................................................................ 35

6.2 Implications from Pipeline Disruption ............................................................. 46

6.3 Implications from Tropical Cyclones ................................................................ 51

Appendices..................................................................................................................57

Appendix A Data Sources ............................................................................................ 57

Appendix B Weather Analysis .................................................................................... 60

Appendix C Statistical Details .................................................................................... 66

Appendix D Texas Railroad Commission Curtailment Plan .......................... 69

Appendix E Liquidity in Texas Natural Gas Market ......................................... 71

Appendix F December 1983 event simulated for 20112012 ..................... 86

Glossary ................................................................................................................................. 99


3/107


BLACK&VEATCH|TableofContents ii

LISTOFTABLES

Table 1 Sources of curtailment information. ................................................................. 12

Table 2 Indicative numbers of natural gas compressors serving the

Texas gas pipeline infrastructure................................................................. 22

Table 3. Sizes of curtailmentincident data sets available for

definition of events and probabilistic risk analyses. ............................ 28

Table 4. Weather stations used for freezingweather analyses of

ERCOT. ..................................................................................................................... 35

Table 5. Winter (Dec, Jan, Feb) daily HDD correlations among the

four heavyload Weather Zones. ................................................................... 35

Table 6. Winter (Dec, Jan, Feb) daily HDD correlations between

ERCOT and other gasdemand regions. ..................................................... 36


4/107


BLACK&VEATCH|TableofContents iii

LISTOFFIGURES

Figure 1 Black & Veatch Approach to Delivery of Phase1 Study ............................. 9

Figure 2 Numbers of documented gascurtailment incidents studied. ................ 13

Figure 3 Fishbone diagram for possible freezingweather causes of

gas curtailments. ................................................................................................. 14

Figure 4 Fishbone diagram for possible pipelinerelated causes of

gas curtailments. ................................................................................................. 15

Figure 5 Fishbone diagram for possible tropicalcyclonerelated

causes of gas curtailments. ............................................................................. 16

Figure 6 Interstate Natural Gas Pipelines Serving ERCOT Generators ............... 19

Figure 7 Intrastate Natural Gas Pipelines Serving ERCOT Generators ................ 20

Figure 8 Natural Gas Storage Assets in ERCOTs Service Region ............................ 21

Figure 9 Natural Gas Pipelines Serving ERCOT Electric Generators ..................... 23

Figure 10 Number of Pipeline Interconnects for Each Electric

Generator ............................................................................................................... 24

Figure 11 Pipeline Capacity as Percentage of Peak Needs ........................................ 25

Figure 12 SupplyDemand Fundamentals. ....................................................................... 30

Figure 13 Black & Veatch Integrated Market Modeling Process............................. 31

Figure 14 Lower48 Natural Gas Supply Projection .................................................. 32

Figure 15 Texas Natural Gas Supply Projections ........................................................ 32

Figure 16 Lower48 Natural Gas Demand Projection .............................................. 33

Figure 17 Texas Natural Gas Demand Projection ....................................................... 34

Figure 18 PowerOutage Risk curve (freezing weather) annualized,

including error envelope beginning at the 55th percentileof probability (mode or peak) of the risk distribution. ....................... 38

Figure 19. Loss of onshore gas production during extreme freezing

events. ...................................................................................................................... 39

Figure 20 PowerOutage Risk curves for freezing weather with

projected future trends .................................................................................... 40

Figure 21. Incremental Daily Demand By Scenario. .................................................... 42

Figure 22. Onshore Gulf Coast Production and Impacts from

Wellhead Freezeoffs. ........................................................................................ 43

Figure 24. Projected Intrastate Pipeline Utilization, North Texas to

Houston. .................................................................................................................. 44

Figure 25. Projected Intrastate Pipeline Utilization, West Texas to

North Texas. .......................................................................................................... 45

Figure 26. Projected Intrastate Pipeline Utilization, South Texas to

Houston. .................................................................................................................. 45

Figure 27. Definition of the pipeline incidents available for risk

analysis. ................................................................................................................... 46


5/107


BLACK&VEATCH|TableofContents iv

Figure 28. PowerOutage Risk curve for pipeline infrastructure

issues. ....................................................................................................................... 47

Figure 29 Pipeline Capacity Comparison Across Pipeline Disruption

Scenarios ................................................................................................................ 50

Figure 30. Indicative Pipeline Utilization Across Pipeline Disruption

Scenarios. ............................................................................................................... 51

Figure 31 Observed frequency of tropical cyclones and their impacts

on gas production in the Gulf of Mexico .................................................... 52

Figure 32. PowerOutage Risk curves derived for annualized tropical

cyclone frequencies. ......................................................................................... 53

Figure 33. Temporary gas demand destruction caused by Hurricane

Ike in 2008. Data from Energy Information

Administration. .................................................................................................... 54

Figure 34. Gulf of Mexico Gas Production Affected by Tropical

Cyclones. ................................................................................................................. 55

Figure 35. Loss of Gas Production Anticipated for Tropical Cyclone

as the Causal Events. .......................................................................................... 56

Figure A1: Process used to collect and sort information about gas

curtailments. ......................................................................................................... 57

Figure E1: Texas Natural Gas Pricing Points. .................................................................. 78

Figure E2: Criteria for Platts Tier Rankings for Natural Gas Pricing

Points. ...................................................................................................................... 79

Figure E3: Historical Platts Tier Rankings for Texas Pricing Points ..................... 79

Figure E4: Traded Volumes and Deals of Natural Gas Reported at

Texas Pricing Points: Monthly Averages of Daily Volumes. .............. 80Figure E5: Daily Traded Volumes Reported at Texas Natural Gas

Pricing Points: September 2001. .................................................................. 81

Figure E6: Daily Traded Volumes Reported at Texas Natural Gas

Pricing Points: February 2003. ..................................................................... 82

Figure E7: Daily Traded Volumes Reported at Texas Natural Gas

Pricing Points: August through October 2005. ....................................... 83

Figure E8: Daily Traded Volumes and Deals Reported at Texas

Natural Gas Pricing Points: September 2008 .......................................... 84

Figure E9: Daily Traded Volumes and Deals Reported at Texas

Natural Gas Pricing Points: January through March 2011. ................ 85

Figure F1. Notably cold winters affecting ERCOT since 1950................................. 86

Figure F2. Lengths of freezingweather events affecting ERCOT. ......................... 87

Figure F3. Daily temperatures in north Texas during December 1983 .............. 88

Figure F4. Freezing temperature patterns in major historical events

affecting ERCOT. .................................................................................................. 88


6/107


BLACK&VEATCH|TableofContents v

Figure F5. Weather stations used for the December 1983 analyses

across ERCOT........................................................................................................ 89

Figure F6. Freezeoff risks at an onshore natural gas wellhead. ........................... 91

Figure F7. Significance of physical temperature relative to wind chill. .............. 92

Figure F8. Daily wind chill values across ERCOT during December

1983. ......................................................................................................................... 92

Figure F9. Natural gas production losses during the February 2011

freezingweather event. ................................................................................... 93

Figure F10. Empirical productionloss models based on production

weather data regressions. ............................................................................... 94

Figure F11. Theoretical gasproduction losses in Feb 2012 under

Dec 1983 weather conditions. ....................................................................... 95

Figure F12. Daily winter freeze risks across ERCOT compared with

December 1983 event. ...................................................................................... 96

Figure F13. Daily winter wind risks across ERCOT compared withDecember 1983 event. ...................................................................................... 97

Figure F14. Estimated Aggregate Loss of Generation Capacity .............................. 99


7/107


BLACK&VEATCH|1.0ExecutiveSummary 3

1.0ExecutiveSummary

This study presents the risk of gas supply curtailment to electric generators within the

service region of the Electric Reliability Council of Texas (ERCOT) over a 1Year, 5Year and

10Year time horizon. It reviews historical incidents of gas supply curtailment experienced

by ERCOTs generators, examines the natural gas infrastructure serving these generators

and assesses the risk of gas supply curtailment on a probabilistic and a fundamental

supply/demand basis.

Curtailment was defined for the purposes of this study as the loss of normally expected gas

delivery as a consequence of supply or transportation interruptions caused by weather

driven, contractual or operational issues.

This study considers the physical capabilities of the natural gas infrastructure in serving

electric generators rather than the contractual arrangements to serve electric generators

with natural gas. Although there may be financial implications to procuring the gas supply

needed, natural gas service is generally available to electric generators subject to the

regulatory and physical constraints of the system. Further, studying contractual

agreements which are subject to commercial negotiations and change through time, does

not allow for a longer term view of the risk of natural gas curtailment to electric generators

which is better captured from the perspective of the physical limitations of the natural gas

infrastructure in serving the needs of electric generators.

This study does not include or consider mitigating measures that have been or can be

incorporated to reduce the risk of gas supply interruption for power generators. Therefore,

this study takes a conservative view on the risk of gas curtailment to electric generators.There have been significant changes in the gas industry over the last 25 years, specifically,

the pipelines typically no longer own the gas they transport and deliver, and there is an

increased use of gas storage as a physical hedge against both supply and pricing volatility

and to ensure deliverability. Combined with the greater liquidity in the natural gas market,

in reality, when natural gas supply or delivery is impacted, the redundancy and

interconnectedness in the natural gas market generally provides consumers (including

electric generators) with alternate sources and routes for natural gas supply to partially or

fully serve their needs. Pipeline linepack, natural gas storage and displacement of supply

from other markets could all contribute to mitigate the risk of disruption of natural gas

supply to electric generators within the ERCOT service region that are presented in this

study.


8/107



Historicalcurtailmentsdatarecordkeepingislimited

This study examined historical

records for gas supply curtailment

from various sources including

ERCOT, the National EnergyTechnology Laboratory (NETL) and

the Railroad Commission of Texas

(TRRC) and found severe limitations

in capturing information about

incidents of natural gas curtailment

to electric generators. The leading

cause of the gas supply curtailment

incidents identified was freezing

weather with existing TRRC

regulations and/or pipeline contractual provisions contributing to gas supply curtailments

to electric generators. Pipeline disruptions and tropical cyclones were inferred to have

caused the other historical incidents of curtailment that were reviewed.

ERCOTgeneratorsdemonstratereliabilityandredundancyofnaturalgassupply

This study conducted a survey of

electric generators within ERCOTs

service region to assess their access to

natural gas infrastructure to serve

their gas demand. Based on survey

responses, ERCOTs electric

generators demonstrate reliability

and redundancy of supply through

their interconnections with multiple

pipelines and access to a level of

capacity that is well in excess of their

peak natural gas needs. 60% of

survey respondents (corresponding to 51,550 MW of nameplate capacity1)indicated

interconnects with more than one natural gas pipeline. All the survey respondents that

provided sufficient data to make an assessment of adequacy indicated access to capacity in

excess of their peak needs.

Naturalgaspipelineinfrastructureissufficienttomeetprojectedneeds

Natural gas pipeline infrastructure serving ERCOT generators was found to be adequate to

meet anticipated peak demand during the analysis period in the scenarios analyzed.

1 The nameplate capacity is inclusive of generation capacity that is part of Private Use Networks which

generally serve their own industrial loads rather than selling power into ERCOT.

108

10

2

54

0

20

40

60

80

100

120

GASCURTAILMENT/

FREEZINGWEATHER

GASCURTAILMENT/PIPELINE

OPERATION

GASCURTAILMENT/

TROPICALCYCLONE

GASCURTAILMENT/

UNKNOWNCAUSE

NumberofIncidents

GasCurtailmentIncidents&Causes

23

11

13

4

65

7

0

5

10

15

20

25

100%149% 150%199% 200%249% 250%299% 300%349% 350%399% >400%

NumberofGenerators

Capacityas%ofPeakDemand

HistogramShowingPipelineCapacityas%ofPeakNeeds


9/107



Although there is potential for isolated incidents, the fundamental supply/demand analysis

undertaken in the study indicated the robustness of the natural gas pipeline infrastructure

in meeting the needs of electric generators within ERCOT, even in the presence of strong

competing demand from other markets and sectors.

Risk

from

Freezing

Weather-

18%

probability

of

having

2000

MW

of

capacity

temporarilyunavailableduetogascurtailments

Risk assessment, based on

historical incidents of

curtailment, indicates that in

any given winter, there is an

18% probability of supply

disruption from lack of gas

supply or

contractual/regulatory defined

curtailment impacting about2,000 MW generation capacity

and about 90% probability of

impacting about 350 MW.

While freezing weather is the

most impactful of the risk

factors considered, the

probability of gas supply curtailment due to freezing weather projected forward should be

viewed together with associated mitigations namely, increased thermal protection of

wellheads against freezeoffs and the priorities and revision of contractual curtailments

initiated by freezing weather. Both wellhead thermal protection and alternative contractual

provisions offer opportunities for assuring greater reliability of gas deliveries during coldwinter events.

RiskfromPipelineDisruptions - 5%probabilityinthenear-termofhaving500MW

ofgenerationcapacitytemporarilyunavailable

Risk assessment of pipeline

disruptions based on historical

incidents of curtailment

indicates that there is a 5%

annual probability of losing

500 MW as a consequence ofgas supply curtailment due to

pipeline outages. Although risk

assessment assumes that the

entire anount of curtailed gas

was required for power

generation and no alternate


10/107



supply was available, redundancy in pipeline capacity serving generators can reduce

exposure to gas supply curtailment from pipeline disruptions.

RiskfromTropicalCyclones

- 13%probabilityinthe

near-

term

of

having

1000

MWofgenerationcapacity

temporarilityunavailable

Compared with the total

volume of gas required for

ERCOT power generation, the

proportion of gas obtained

from Gulf of Mexico (GOM)

offshore production is small

with less than 5% of the total

ERCOT gas consumptiondepending on GOM

production. Therefore, tropical cyclone impacts on ERCOTs power generation are

relatively small. For perspective, on the 1Yr horizon there is a 13% risk of 1000 MW

generation loss from tropical cyclone.

Conclusions&recommendations

Data availability placed constraints on understanding and analyzing historical gas supply

curtailments to electric generators within ERCOTs service region. Increased coordination

between natural gas and power industry regulating agencies could help ensure improved

crosscapture of information as the role of natural gas as a fuel source for power generation

continues to grow. If ERCOT is expected to monitor fuel impacts on the reliability of the

electric grid, better data capture of curtailment incidents is needed.

Some specific recommendations are listed below:

ERCOT Operator logs were the most complete source reviewed in the study of incidentdata on natural gas supply disruption experienced by electric generators within ERCOT.

It was observed that capture of natural gas curtailment incident information would be

more complete and accurate with greater training of ERCOT operators to improve

recognition of, and familiarity with, natural gas pipelines and utilities serving ERCOTs

electric generators.

This study included a survey of gasfired electric generators within ERCOTs serviceregion to assess their experience with natural gas supply disruption. It is recommended

that standardized categories of gas delivery issues should be included as a regular report

element in the annual reporting by generators to ERCOT. This will allow ERCOT to track

and assess any trends associated with natural gas supply disruption to electric generators

and to develop risk mitigation plans if a trend reflecting increasing disruption to electric

generators is observed.


11/107



This study recommends continued coordination between ERCOT and the RailroadCommission of Texas (TRRC) to facilitate better data capture including development of

communication pathways and reports for gasdelivery incidents affecting power

generation facilities.

In addition to cost considerations associated with the decision to contract for firm orinterruptible gas service and /or have duel fuel supply, contractual agreements that

require curtailment of gas supply to generators or mandatory curtailment policies as

defined by the TRRC may inhibit a power generators ability and motivation to acquire

firm gas supply. Review of these agreements and policies could help determine whether

new policies or regulations are required to increase the reliability of ERCOT generation.


12/107


BLACK&VEATCH|2.0Introduction 8

2.0IntroductionThe Electric Reliability Council of Texas (ERCOT) commissioned a Gas Curtailment Risk

Study to evaluate the risk of natural gas supply disruptions to electric generating stations

within the ERCOT administered portion of Texas.

During the first week of February 2011, the Southwest experienced extremely cold weatherwith temperatures falling by as much as 50 degrees over an eighteenhour period in various

cities in Texas. This extreme cold event saw low temperatures in the DallasFt. Worth area

dipping to 13 F which, according to our probabilistic analysis, was an event with a winter

time daily probability of less than 1%. During the first four days of February, 210 individual

generating units within ERCOTs service region experienced disruption of their normal

generation operations due to a variety of factors. The scale of generation loss led to

controlled load shedding that impacted as many as 4.4 million customers2 during the event.

A majority of the generation losses experienced occurred due to problems related to plant

operation including frozen sensing lines, frozen equipment, frozen water lines, frozen

valves, and blade icing. Extreme low temperature events in 1989 and 2003 similarly

created conditions resulting in loss of generation in ERCOT. A FERCNERC investigationfound that, although the generation loss associated with these extreme weather events was

not primarily driven by gas supply curtailment3, natural gas supply was impacted as a result

of weather and contributed to the loss of generation.

By fuel type, about 38% of ERCOTs annual average generation is currently accomplished

with natural gas.4 Gasfired generation capacity within ERCOT is projected to increase by

over 15,000 MW in the next 10 years. With natural gas share of electric generation within

ERCOT being poised to increase to 50% over the next 10 years and beyond, it is important

to understand the risks faced by electric generators due to potential disruptions in natural

gas supply.

This study is intended to increase ERCOTs understanding of the risks of generation loss

from gas supply curtailment in the future and to consider potential mitigation measures

that ERCOT can pursue to reduce risks arising from these curtailments. The study is also

intended to assist ERCOT to objectively assess the costs and benefits of planning operations

for mitigating gas supply curtailment risk to its electric generators.

The scope covered by this study is summarized below:

2Report on Outages and Curtailments During the Southwest Cold Weather Event of February 1-5, 2011,Federal Energy Regulatory Commission (FERC) and North American Electric Reliability Coporation

(NERC), August 2011.3 It was stated that For the Southwest as a whole, 67 percent of the generator failures (by MWh) were due

directly to weather-related causes, including frozen sensing lines, frozen equipment, frozen water lines,

frozen valves, blade icing, low temperature cutoff limits, and the like. (p. 8). Report on Outages and

Curtailments During the Southwest Cold Weather Event of February 1-5, 2011, Federal Energy Regulatory

Commission (FERC) and North American electric Reliability Coporation (NERC), August 2011, 357 p.4 ERCOT (2011b) Data file: GenerationByFuelType_2002-2010.xls.http://www.ercot.com/content/news/presentations/2011/GenerationByFuelType_20022010.xls.


13/107



Deliverable 1 Review past natural gas interruptions impacting electric generation forinsights.

Deliverable 2 For pipeline systems that serve generation, survey pipeline physicallimitations to providing natural gas to electric generation in ERCOT.

Deliverable 3 Review scenarios in which ERCOT natural gas supply to electricgenerating stations could be significantly limited, including conditions of severe cold

temperature combined with high wind speeds. Calculate the risk (assess probability) of

such events in the near (1 to 5 year) and mid (5 to 10 year) timeframe.

Black & Veatchs approach to meeting the requirements of the three deliverables was

designed to collect, process and systematically analyze the data required to estimate the

risk of gas supply curtailment to the electric generators within ERCOTs service region and

is illustrated in Figure 1.

Figure1Black&VeatchApproachtoDeliveryofPhase1Study

2.1ORGANIZATIONOFTHISREPORT

The remainder of this report is organized as follows:

Section3: Review of Historical Curtailment Summary of our review of the historically

reported incidents of natural gas curtailments within ERCOT.

Section4: Natural Gas Infrastructure Summary of the natural gas infrastructure serving

electric generators within ERCOT s service region.

1. CompilePastNatural

Gas

Interruptions

for

PowerGeneration

A. Events

(numbers&types)

B. CausalFactors

C. LessonsLearned

2. SurveyGasPipeline

Data&Performance

A. Transmission

B. LDCs

C. Storage

MapoverofPipelinestoGasFiredGenerators

ReferenceDatabaseofRealizedRisksandConsequences

3. ConstructGas

CurtailmentScenarios

A. ExogenousRisksB. ProbabilisticRisk

Analyses:5 and10yrHorizons

C. ErrorEstimationsforProbabilisticRiskAnalyses

IdentificationofScenarios

S ev er eWeather Infrastructure

Disruptions

ProbabilisticAnalysisofScenarios

PalisadeDecisionToolsmodeling

AssessingImpactonNaturalGasService

ModelingwithGPCM

D. ERCOTSpecificRiskedCurtailments


14/107



Section5: Risk Assessment Approach & Assumptions Overview of overall approach and

analytical tools and a list of key assumptions underlying Black & Veatchs analysis.

Section6:Risk Assessment Results Discussion on analytical approach, scenarios

examined and the results of risk assessment

Finally, we include Appendices that provide more detailed descriptions, information andresults from the study.


15/107


BLACK&VEATCH|3.0ReviewofHistoricalCurtailments 11

3.0ReviewofHistoricalCurtailmentsAn understanding of ERCOTs historical experience with natural gas curtailments is an

important first step while examining the risk of any potential future disruptions to natural

gas supply to electric generators within ERCOTs service region. We undertook a review of

available historical data on natural gas curtailment incidents in order to collate and examine

the experience to date with natural gas curtailment to electric generators.

It should be noted that the term curtailment has different definitions depending on the

industry and the agency that utilizes it. The Federal Energy Regulatory Commission (FERC),

National Energy Technology Laboratory (NETL) and the Railroad Commission of Texas

(TRRC) each has a different definition and understanding of the term curtailment as it is

applied within their jurisdictions. For the purpose of this study, the working definition for

curtailment that is Lossofnormallyexpectedgasdeliveryasaconsequenceofsupplyor

transportationinterruptionscausedbyweather-driven,contractualoroperationalissues.

Black & Veatch conducted research, using publiclyavailable information sources, to gather

facts about historical cases of natural gas delivery interruptions within Texas that haveimpacted gasfired electric power generation. It should be noted that gas supply

interruptions also can occur due to contractual provisions, TRRC defined regulations

requiring disruptions of gas supply to power generators, as well as nondelivery of

contracted supply5.

Black & Veatch also worked with ERCOT to locate event data that is relevant to natural gas

supply reliability. Historical ERCOT Monthly Operations Reports6, NERC System

Disturbance Reports7, various FERC issued reports and historical pipeline operational

information were sources of timeline information. In addition, ERCOT issued a survey

questionnaire prepared by Black & Veatch to the natural gasfired electric generators within

its service region seeking information on natural gas curtailments experienced by thegenerators during their operational history. The information gathered from these various

sources were reviewed to compile chronological timelines for events involving curtailments

or other disturbances of natural gas supplies to generation facilities. Although the focus of

the study was on the ERCOT region, gasrelated incidents elsewhere were reviewed to the

extent that they offer insights into issues relevant to ERCOT.

For each occurrence of natural gas interruption that was identified, Black & Veatch

examined the causal factors leading to the gas interruption. It should be noted that gaps in

data availability and historical recordkeeping placed constraints on examining and

accurately determining the cause of every incident of curtailment that was reviewed.

Causal factors that were investigated include:

Severe cold weather conditions in ERCOTSevere cold weather conditions in regions of competing gas demand5 TRRC Gas Curtailment Plan of 1973. Oil and Gas Docket, Gas Utilities Division, No. 20-62, 505,

Docket No. 489, January 5, 1973. http://www.rrc.state.tx.us/meetings/dockets/docket489.php6 ERCOT Operations Monthly Reports. http://www.ercot.com/mktinfo/reports/omr/7 NERC System Disturbance Reports. http://www.nerc.com/page.php?cid=5|66


16/107



Operational interruptions caused by pipeline outagesTropical cyclones (hurricanes, tropical storms and tropical depressions)OtherBlack & Veatch examined the common threads linking the instances of natural gas

interruption identified and their causes to identify lessons that can be learned by ERCOT

from these historic experiences.

3.1DATAAVAILABILITY&SOURCES

Historical records of natural gas curtailment to electric generators were found during our

review to be limited. Our review found that most of the data available was for the last

decade rather than for previous time periods, reflecting better record keeping in more

recent years. The primary data sources that were examined as potential sources of records

of historical natural gas curtailment are summarized in Table 1.

Table1 Sourcesofcurtailmentinformation.

SOURCENAME

AVAILABLE

STARTING

DATE

AVAILABLE

ENDINGDATE NOTE

ERCOTGasCurtailmentSurvey 9/13/1994

(Earliest

Curtailment

Reported)

4/1/2011(Latest

Curtailment

Reported)

Responsestosurvey

senttonaturalgas

firedelectric

generatorswithin

ERCOT'sserviceregion

aspartofthisstudy

ERCOTOperatorLogs Dec2002 Aug2011 OperatorLogs

providedbyERCOT

filteredusingthekey

wordsGas

CurtailmentandGas

Restriction

ERCOTMonthlyOperationsReports Jan2004 Jul2007 FocusedonOperating

ConditionNotice

NationalEnergyTechnologyLaboratory

(NETL)ElectricDisturbanceEvents(OE

417)AnnualSummaries

Year2000 Year2011 FocusedontheMajor

ElectricDisturbances

andUnusual

Occurrences

RailroadCommissionofTexas(TRRC)

PipelineIncident

Reports

and

separate

responsetoERCOTDataRequest

12/12/1983 2/2/2011 TRRCresponsetodata

requestsent

as

part

of

thisstudy

Secondary sources of data that were examined and utilized are listed in Appendix A Data

Sources.


17/107



There was limited overlap between curtailment or disruption data available through

natural gasfocused and powerfocused entities. NETL and other sources of curtailment

data from powerfocused entities placed limited or no emphasis on capturing or reporting

the natural gas fuel aspect of recorded events and, at best, natural gas curtailment could

only be inferred for some of those incidents. Pipeline electronic bulletin boards and other

natural gasfocused sources, in turn, did not capture impacts of gas curtailment events onelectric generators in detail although such impacts were inferred by Black & Veatch where

possible. ERCOTs operator logs were most directly applicable of the various primary data

sources reviewed. Documented gas curtailments outside of contractual agreements were

relatively rare among the incidents reviewed with most curtailment incidents reviewed

appearing to be contractually permitted.

3.2SUMMARYOFHISTORICALCURTAILMENTS&CAUSES

In all, 216 incident records were identified upon review of the various data sources that

were examined. The majority of historical curtailment incidents reported for ERCOT were

winter occurrences associated with freezing weather as shown in Figure 2. A key finding of

those incidnents is that the majority of historical curtailments to electric generators withinERCOTs service region during freezing weather appear to have been contractually

permitted and triggered by a temperature threshold. A small number of coldweather

related incidents were

attributed to physical

disruption of upstream

supply or

infrastructure. A FERC

NERC report8, for

example, attributed a

majority of the

February 2011

generation loss to

problems with

winterization related to

plant operations and

with a smaller portion

attributed to gas supply

loss from wellhead

freezeoffs and field

level infrastructure

failures. Figure 3 shows a fishbone diagram9

outling possible causes and effects leading togas system failure related to freezing weather. In a failure modes and effects analysis

8Report on Outages and Curtailments During the Southwest Cold Weather Event of February 1-5, 2011,

Federal Energy Regulatory Commission (FERC) and North American Electric Reliability Corporation

(NERC), August 2011, 357 p.9 A fishbone diagram (also known as an Ishikawa diagram) is a tool used to identify failure pathways in a

failure mode and effects analysis (FMEA). In the current study, fishbone diagrams are used to summarize

108

10

2

54

0

20

40

60

80

100

120

GASCURTAILMENT/

FREEZINGWEATHER

GASCURTAILMENT/PIPELINE

OPERATION

GASCURTAILMENT/

TROPICALCYCLONE

GASCURTAILMENT/

UNKNOWNCAUSE

NumberofIncidents

GasCurtailmentIncidents&Causes

Figure2Numbersofdocumentedgascurtailmentincidentsstudied.


18/107



(FMEA), these are possible causeandeffect strings that can affect gassystem performance,

based on general historical experience. The precise causeandeffect string is not always

expressly published for every curtailment event. The potential factors leading to gas supply

disruptions due to freezing weather are 1) freezing of onshore gas wellheads, 2) onshore

power grids trip and pipelines lose pressure as gas compressors and/or Supervisory

Control and Data Acquisition (SCADA) systems lose power and 3) contractual provisionswith gas suppliers/transporters that allow curtailment of gas supply to power generators

based on temperature thresholds.

Figure3 Fishbonediagramforpossiblefreezingweathercausesofgascurtailments.

Pipeline operations represented the next largest driver of natural gas curtailment incidents

historically. Those incidents were caused by unscheduled maintenance and line ruptures.

There were 10 reported incidents of gas curtailments related to pipeline disruptions in the

data reviewed. In addition, 54 incidents of gas curtailment were reported without any

specified cause although our further research showed that none were linked either to

freezing temperatures in winter or tropical cyclone occurrences in summer. Since those

how causative agents might lead to gas curtailments but without identifying likelihood of the alternative

pathways.


19/107



incidents were not weatherrelated, they were assumed to be infrastructurerelated and

grouped together with pipeline disruptions. Figure 4 shows a fish bone diagram examining

the cause and effect leading to gas system failure related to pipeline disruptions.

Figure4 Fishbonediagramforpossiblepipelinerelatedcausesofgascurtailments.

Only 2 incidents of gas curtailments driven by tropical cyclones were observed in the

reviewed data10. Figure 5 shows a fishbone diagram examining the possible causes and

effects leading to gas system failure related to tropical cyclones. The three main failurepaths driven by tropical cyclones are 1) shutting of offshore platforms due to a storm in the

Gulf of Mexico (GOM); 2) onshore flooding caused by excessive rainfall that impacts gas

processing facilities; and 3) high winds associated with tropical storms knock down power

lines and cutoff power to gas pipeline compressors and/or SCADA systems.

10 Weather-related incidents in the ERCOT Operator Logs dated from 2002 and later. Major tropical

cyclone landfalls and coastal flooding events occurred in 1989 and 2001 prior to first records in the

ERCOT Operator Logs.


20/107



Figure5 Fishbonediagramforpossibletropicalcyclonerelatedcausesofgascurtailments.

3.3BESTPRACTICES&LESSONSTOBELEARNED

A relatively small number of curtailment incidents outside of contractual agreements were

observed overall in the data reviewed as part of this study. This would indicate that natural

gas supply has proven to be a reliable fuel source for power generators operating in ERCOT

and that market liquidity and commercial agreements appear to largely be effective in

procuring natural gas supply for electric generators. The growth of onshore unconventionalnatural gas resources in Texas may be expected to help make natural gas supply even more

readily available for ERCOT generators.

Survey responses indicate that some electric generators in the DallasFort Worth region

have entered into contractual agreements that allow curtailment of their natural gas supply

in the event of extreme cold weather which is driven in part by curtailment priorities

defined by the TRRC. Appendix D of the report includes the Curtailment Plan requirements

of the TRRC in more detail. In addition to regulatory requirements, contractual agreements

also can reflect a tradeoff between the cost of firm supply and the costs for contractual

interruption based on historical experience that natural gas supply is available when

needed during most days of operation.

Switching to oil was observed in historical data as a mitigation measure when gas

curtailments were in effect due to contractual terms. It should be noted that the economics

of switching may place restrictions on the ability to switch to oil going forward.


21/107



Connectivity to multiple pipelines or to storage facilities would provide both supply

flexibility to minimize delivered gas supply costs and fuelsupply redundancy for generators

when curtailed by one pipeline.

Increased coordination between natural gas and power industry regulating agencies could

help ensure crosscapture of information as the role of natural gas as a fuel source forpower generation continues to grow. If ERCOT is expected to monitor fuel impacts on the

reliability of the electric grid, better data capture of curtailment incidents is needed. This

study recommends the following measures to better capture information related to natural

gas supply and curtailments to electric generators:

Training of ERCOT operators to improve their familiarity with the natural gasinfrastructure will help to increase the data accuracy of operator logs. As identified in this

study, the operator logs comprised the most compete data source recording the

disruption of natural gas supply to electric generators within the ERCOT service region.

Adding standardized questions on gas deliveryrelated issues into annual reportssubmitted by the electric generators will allow ERCOT to track and assess any trends

associated with natural gas supply disruptions that electric generators experience. It will

also help ERCOT to understand and manage gas supplyrelated risks.

Continued and growing coordination between ERCOT and the TRRC in development ofreports for capturing and sharing information on gas supply delivery issues impacting

electric generators. Coordinated actions can foster better data capture for both

organizations.

Review of gassupply agreements and gascurtailment policies could help determinewhether new policies or regulations are required to increase the reliability of ERCOT

generation. In addition to cost considerations, contractual agreements that require

curtailment of gas supply to generators or mandatory curtailment policies as defined by

the TRRC may inhibit a power generators ability and motivation to acquire firm gas

supply.


22/107


BLACK&VEATCH|4.0NaturalGasInfrastructure&Market 18

4.0NaturalGasInfrastructure&MarketTexas is the largest producer as well as consumer of natural gas in the U.S., contributing

about onethird of the total production in the U.S. and consuming oneseventh (with over

85% of it being consumed in the industrial and electricity generation sectors)11. As a

consequence, Texas enjoys one of the most robust natural gas markets in North America

with welldeveloped infrastructure that includes natural gas production facilities, natural

gas processing facilities, interstate and intrastate natural gas pipelines and natural gas

storage facilities.

Texas leads all states in the U.S. in the number of pipeline miles with more than 21,000

miles of interstate natural gas transmission pipelines and more than 130,000 miles of

intrastate natural gas transmission and distribution pipelines making it one of the best

connected and served markets in the U.S.12

4.1 INTERSTATEPIPELINES

Among the major interstate pipelines serving electric generators within ERCOTs service

region are Texas Eastern Transmission, CenterPoint Energy, El Paso Natural Gas, NaturalGas Pipeline Company of America, Tennessee Gas Pipeline, and Transcontinental Pipeline

(Figure 6). Most of these interstate pipelines transport production from the Gulf Coast

region and flow north to serve the Midwest market and northeast to serve the East Coast

markets of the U.S. El Paso Natural Gas moves gas produced in West Texas fields to serve

the West Coast.

Since those pipelines move gas that is produced in and near Texas to consumers in the

Midwest, East Coast and West Coast, the other market destinations can be considered as

representing competition for natural gas supply for electric generators within ERCOTs

service region. The potential for gas supply disruption to electric generators within

ERCOTs service region that is posed from competing demand served by these pipelines isone of the risk factors considered in this study and is discussed in detail in Section 5 and

Section 6.

11 Natural Gas Annual Supply & Disposition by State, US Energy Information Administration.

http://www.eia.gov/dnav/ng/ng_sum_snd_dcu_nus_a.htm12 Texas Pipeline System Mileage, updated October 28, 2010,, Railroad Commission of Texas.

http://www.rrc.state.tx.us/data/gasservices/vitalstats/mileage.php


23/107



Figure6

InterstateNatural

Gas

Pipelines

Serving

ERCOT

Generators

4.2 INTRASTATEPIPELINES

Texas holds the distinction of having the largest number as well as the most miles of

intrastate pipelines in the U.S. Those pipelines gather and transport natural gas from

supply basins in Texas to local gas distribution companies, electric generation and

industrial and municipal consumers, as well as to connections with intrastate pipelines and

interstate pipelines that transport this gas to enduse markets in the Midwest, East Coast

and West Coast. The major players in the intrastate pipeline market include Atmos Energy

Corporation, Enterprise Products Partners, L.P., Kinder Morgan Energy Partners, L.P. and

Energy Transfer Partners, L.P. owning and operating multiple, large Texas intrastate

pipelines between them (Figure 7). Texas intrastate pipelines are regulated by the TRRCand are subject to alternative regulations compared with those of FERCregulated

interstate pipelines. The TRRCs oversight over intrastate pipelines is largely focused on

safety and pipeline integrity with less oversight, when compared to FERCregulated

pipelines, related to commercial issues. This can result in less transparency about the

available capacity and the transportation costs associated with intrastate pipelines when

compared to interstate pipelines. It should be noted, however, that the intrastate pipeline


24/107



market in Texas is highly competive when multiple pipeline or supply alternatives are

available to an enduser.

Figure7IntrastateNaturalGasPipelinesServingERCOTGenerators

4.3 NATURALGASSTORAGEFACILITIES

Natural gas is commonly stored in underground rock formations such as depleted oil and

gas reservoirs or leached caverns in salt domes. Natural gas storage helps to match the

relatively constant production profile of natural gas with its highly seasonal consumption

pattern by creating flexibility in the market and allowing participants to storage large

volumes of natural gas in summer when the traditional heating load is typically low and to

use this stored gas in winter when the heating load increases. Texas stands fourth in theU.S. in total underground natural gas storage capacity with over 783 Bcf of storage

capacity13. Figure 8 shows the underground natural gas storage assets within ERCOTs

service region.

13 Underground Natural Gas Storage Capacity, as of December 2010, updated December 29, 2011, US

Energy Information Administration.

http://www.eia.gov/dnav/ng/ng_stor_cap_a_EPG0_SAC_Mmcf_a.htm


25/107



In addition to helping balance the seasonal demand with relatively constant production,

natural gas storage also offers shortterm flexibility to the natural gas market by being a

source of supply when demand is higher than anticipated and being able to absorb supply

when demand is lower than anticipated. It is this attribute of natural gas storage that makes

it attractive to electric generators seeking to manage the daytoday volatility in their gas

supply needs. Natural gas storage offers electric generators the ability to quickly accesssupply when their generation needs ramp up or an alternate destination for surplus natural

gas supply when generation needs ramp down. High deliverability storage or storage with

the ability to inject or withdraw high volumes of gas each day relative to the total storage

capacity of the field offers the most flexibility to swing with the daily gas supply needs of

electric generators.

Storage assets in Texas include both regulated assets that are part of natural gas pipeline

systems as well as standalone storage assets managed by independent operators.

Figure8NaturalGasStorageAssetsinERCOTsServiceRegion

4.4 ROLEOFGASCOMPRESSORS

Gas production fields, storage fields and both intrastate and interstate pipelines depend

upon gascompression technologies to sustain their operations. A significant loss of


26/107



compression can interrupt pipeline flows and threaten curtailment of gas deliveries to

customers.

Gas compressors are built on either reciprocating or centrifugal technologies and with

either combustionpowered or electricpowereddrive driver technologies. Combustion

driven compression historically has used gas provided by a pipeline and thereby has offered

a significant level of selfsufficiency for pipeline operations. In contrast, electricdrive

compressors depend upon electrical power which is purchased from an outside source (not

controlled by the pipeline) which represents a risk factor beyond the control of the pipeline.

Based upon Black & Veatch research, Table 3 summarizes the proportions of gas and

electricdrive compressors installed in Texas. Although the numbers in Table 3 are not

represented as a comprehensive inventory, the most significant message is that about 18%

of all transmission pipeline compressors are electricdrive and therefore at risk to power

outages. However, pipeline operations often are designed to be able to continue with

limited compressor outages whether gas fired or electricdrive.

Table2IndicativenumbersofnaturalgascompressorsservingtheTexasgaspipelineinfrastructure.

COMPRESSOR

TYPEGASFIELD

TRANSMISSION

PIPELINE

UNDERGROUND

STORAGETOTAL

GasCombustion 4 498 34 536

Electric 1 111 2 114

Total 5 609 36 650

Since the 1990s, the general trend among gas pipelines has been toward selection of

electricdrive compressors based on benefits of lower maintenance costs, lower noise

emissions and lower air emissions14. Any corresponding increase in risks of power outagesgenerally has been considered an acceptable tradeoff.

4.5 SURVEYRESULTSOFNATURALGASINFRASTRUCTURESERVINGERCOTGENERATORS

As part of this study, Black & Veatch conducted a survey, through ERCOT, of the natural gas

fired electric generators within ERCOTs service region to assess the natural gas

infrastructure serving their facilities. The survey requested information on the pipelines,

local distribution companies (LDCs) and storage facilities serving each electric generator.

The information provided through survey responses was supplemented by a number of

data sources to create a compilation of the natural gas infrastructure serving electric

generators within ERCOTs service region. The data sources utilized include FERC, TRRC,

pipeline electronic bulletin boards (EBBs), the US Energy Information Administration (EIA),

Black & Veatchs proprietary database underlying our large body of work in natural gas

market analysis and thirdparty vendor data. While the individual survey results are

confidential, we share the following observations summarizing the survey responses:

14Factors That Influence the Selection of Electric Motor Drives For Natural Gas Compressors, Prepared

for The INGAA Foundation, Inc. by Southwest Research Institute, SwRIProject 18-2090, April 1999, 58 p.


27/107



The survey on natural gas infrastructure was sent to 109 gasfired electric generatorswithin ERCOTs service region and 82% of electric generators fully or partially responded

to the survey. The survey results summarized below are applicable to this population of

respondents alone.

There is diversity in the natural gas infrastructure serving the electric generatorssurveyed with multiple pipelines serving these generators, namely, 44 different natural

gas pipeline systems delvering powergeneration fuel within ERCOTs service region. Of

these 44 pipelines, 7 are interstate natural gas pipelines and 37 are intrastate natural gas

pipelines. Figure 9 shows the top 10 pipelines serving the electric generators in ERCOTs

service region.

Approximately 60% of the generators that responded to the survey (corresponding to51,550 MW of nameplate capacity15) have access to more than one natural gas pipeline

interconnect which can create redundancy in natural gas supply alternatives (Figure 10).

Figure9NaturalGasPipelinesServingERCOTElectricGenerators



0

5

10

15

20

25

30

Nu

mberofGeneratorsServed

NaturalGas

Pipelines

Serving

ERCOT

Electric

Generators


28/107



Figure10NumberofPipelineInterconnectsforEachElectricGenerator

All the generators that provided information on their capacity and peak needs noted thatthey had adequate pipeline capacity to meet their peak demand. Over 65% of these

generators (corresponding to over 39,400 MW of nameplate capacity16) indicated that

they had access to capacity in excess of 150% of their peak needs. Figure 11 shows thelevel of redundancy in pipeline capacity that was reported by the survey respondents. As

seen in this histogram, many electric generators have access to substantial excess

pipeline capacity that can be expected to increase their reliability of supply and offset the

impacts of any supply or pipeline disruptions.

Access to, or contracts for, gas storage appears to be limited, although storage is used on adaily basis by gas suppliers and interstate and intrastate pipelines to manage flows on

their systems. Only 23% of the respondents reported information on natural gas storage

as part of their supply portfolio.



37

27

16

6

4

0

5

10

15

20

25

30

35

40

1 2 3 4 5

NumberofGenerators

NumberofPipelineInterconnects

NumberofPipelineInterconnectsForEachElectricGenerator


29/107



Figure11PipelineCapacityasPercentageofPeakNeeds

Overall, based on the responses by generators to the survey, it appears that ERCOTs

electric generators create reliability and redundancy of gas supply capability through their

interconnections with multiple pipelines and access to a level of capacity that is well in

excess of their peak natural gas needs.

23

11

13

4

65

7

0

5

10

15

20

25

100%149% 150%199% 200%249% 250%299% 300%349% 350%399% >400%

NumberofGenerators

Capacityas%ofPeakDemand

HistogramShowingPipelineCapacityas%ofPeakNeeds


30/107


BLACK&VEATCH|5.0 RiskAssessment Approach&Assumptions 26

5.0 RiskAssessment Approach&Assumptions

5.1 SUMMARYOFAPPROACH

Black & Veatch approached the risk assessment as a combined economic and quantitative

analysis with the final objective being development of riskbased likelihoods of natural gas

curtailments that could affect gasfired generation for scenarios that are specific to ERCOT.

Our work effort for Deliverable 3 was focused on three main analytical efforts:

A. Identification of ScenariosB. Probabilistic Analysis of ScenariosC. Fundamental Analysis of ScenariosA.IdentificationofScenarios

This study utilized a scenariobased approach to assess the risk of gas supply curtailment to

electric generators within ERCOTs service region. Black & Veatchs review of historical

curtailment events provided the basis for selecting and shortlisting potential risk scenariosfor ERCOT to be analyzed in this study. As discussed in Section 3, freezing weather was

found to be the most significant risk factor driving natural gas supply curtailment to electric

generators within ERCOTs service region. Over 60% of the recorded incidents of gas

supply curtailments to electric generators that were reviewed were driven by a freezing

weather occurrence, including some driven by contractual provisions which stipulated

temperature milestones in their curtailment schedules. Other risk factors that were found

during our review of historical curtailments were pipeline disruptions and tropical

cyclones. Accordingly, the scenarios shortlisted for the study are primarily weatherdriven

or infrastructuredriven and listed below:

1. Freezing weather in Texas and outside Texas2. Pipeline disruptions3. Tropical cyclonesFor each given scenario, the study examined a family of occurrences of increasing severity

to facilitate understanding of the shape of the risk profile associated with a given risk factor

as opposed to a point estimate of the risk.

Notable historical curtailments where available were utilized as benchmarks within the

scenarios analyzed.

B.ProbabilisticAnalysisofScenarios

The next analytical step was the probabilistic analysis of each of the scenarios to determinethe risk associated with their occurrence. Those analyses adopted an empirical approach

which distinguished frequencies of potential causal events from frequencies of documented

curtailment events17. Most notably, frequencies of occurrence of problematical weather can

17 A causal event is an environmental or operational factor which, based on historical experience, could

cause a gas curtailment event. For events associated with weather, it is possible to derive causal-event

statistics which are independent of curtailment-event data.


31/107



be calculated for years where records for curtailment events do not exist. But the

information of greatest interest remains in the curtailmentevent reports. Accordingly, the

approach adopted in this study emphasized analysis of curtailmentevent data directly

wherever possible.

The probabilistic (stochastic) methodology employed the following sequence of actions:

Compile curtailment event data for each scenario as defined by a causal relationship (forexample, freezing temperature, tropical cyclone, pipeline failure)

For each scenario, differentiate the curtailment event data into subpopulations ifpossible (for example, curtailments associated with freezing temperatures or high

heatingdegree day numbers either in Texas or elsewhere)

For each population (or subpopulation) of curtailment events in each scenario, employstatisticalanalysis software to derive a bestfit probability distribution function (PDF)

that describes frequency of event occurrence

To the extent that documented reports allowed, derive PDFs

that describe frequency oflost generation by size (MW) in each scenario (This was possible for freezingweather

and pipelineoutage scenarios.)

As necessary, mapover causaleventPDFs onto selected curtailment thresholdsexpressed in units of generation (MW). (This was necessary for tropical cyclones.)

Employ the PDFs,alongwithscalingfactorsforgrowthordeclineofgasdependencies, toderive probabilities of occurrence of the subject hazard (curtailment event) at selected

timeline thresholds (for example, 5 and 10year).

The robustness of probabilistic results for causal events was strongest for weather data

which comprised large, continuous data sets. Probabilistic results for curtailment event

data carried much larger uncertainties associated with the much smaller and lesscontinuous nature of their data. For daily weather data compiled for winter months

(December, January, February) over the period of January 1981 through February 2011, the

data set available for each station typically comprised 2,766 measurements. In contrast,

curtailment event data were typically limited to fewer than 100 incident reports (Table 3).

After incidents were analyzed to define discrete events, and especially as subcategories

were sought among the types of events, the data available to define a subscenario were

reduced to 25 or fewer examples in many cases. For comparison, science and engineering

analyses commonly find that the minimum number of samples required for application of

distributionfunction statistics falls in the range of 155018 which, in the current study, is

matched by the freezingweather and pipelineoutage incident reports but not by the

tropicalcyclone incident reports (Table 3). Accordingly, analyses for freezingweather and

pipelineoutage events proceeded directly using loss reports (typically MW rather than

Bcf/d, based on relative numbers of available reports) but for tropicalcyclone incidents it

18 The minimum sample size depends on the distribution function chosen but the minimum number

generally increases as the function differs from a Normal (Gaussian) distribution. See, for example, Meyer

S. L. (1975)Data Analysis for Scientists and Engineers, John Wiley & Sons, Inc., New York, 413 p..


32/107



was necessary to model indirectly using frequency of occurrence of tropicalcyclone activity

and implied impacts on gas supplies which then were translated to equivalent MW losses in

ERCOT. The effects of different sample sizes on goodnessoffit and on uncertainties (error

bars) for risk levels is illustrated in Appendix B Weather Analysis.

Table3. Sizesofcurtailmentincidentdatasetsavailablefordefinitionofeventsandprobabilisticrisk

analyses.

PERIOD

OF

RECORD

FREEZINGWEATHER(1) TROPICALCYCLONE(2) PIPELINE(3)

Total

Incl.

MW

Loss

Reports

Incl.

Bcf/d

Loss

Reports

Total

Incl.

MW

Loss

Reports

Incl.

Dth/d

Loss

Reports

Total

Incl.

MW

Loss

Reports

Incl.

Dth/d

Loss

Reports

2011

2002106 62 7 2 1 1

2011

198764 46 0

(1)

Excludes

two

other

events

(Dec

1983,

Dec

1989)

for

which

MW

and

Dth/d

loss

data

were

not

available.

Includes

incidentscausedbycontractualprovisions.

(2) Excludestwootherlandfallandfloodingevents(1989,2001)forwhichERCOTrecordsdonotexist.(3) Includes combination of events documented as pipeline plus other events documented as unknown but

whereapipelineinvolvementwasinferred.

C.FundamentalAnalysis

For each scenario identified, Black & Veatch used a fundamental supplydemand model to

examine the sufficiency of the natural gas infrastructure and natural gas supply to impact

on natural gas service to electric generators in the ERCOT region. This analytical approach

incorporates a network representation of the physical capabilities of the natural gas

infrastructure in serving electric generators rather than the contractual obligations on the

natural gas facilities. Although there may be financial implications to procuring the gassupply needed, natural gas service is available to electric generators subject to the

regulatory and physical constraints of the system. In addition, procuring all the bilateral

contracts required to comprehensively capture the contractual obligations within ERCOT is

a significant, if not impossible, undertaking that is complicated by the lack of publicly

available information.

Black & Veatch estimated the natural gas demand or supply implications associated with

each of the scenarios and used these modified demand and supply assumptions as inputs to

the fundamental market model. The analysis examined any resulting constraints within the

system (caused either through increased demand or decreased supply) that impacted the

availability of natural gas supply to electric generators within the ERCOT region.

The fundamental analysis is intended to supplement the probabilistic risk analysis by

defining specific forwardlooking scenarios that examine the sufficiency of pipeline

infrastructure and natural gas supply to meet the needs of electric generators in ERCOTs

service region.


33/107



5.2TOOLSANDSOFTWARE

Black & Veatch utilized a combination of tools to analyze natural gas markets and

infrastructure related to ERCOTs Gas Curtailment Risk Study.

ProbabilisticRiskModeling

Probabilistic risk analyses were performed using the Palisade DecisionTools Professional5.0 software package which includes the module,@Risk, for stochastic simulation through

Monte Carlo and Latin Hypercube algorithms. For each data set analyzed, the empirical data

were passed through the @Risk bestfit functions using AndersonDarling criteria19 to

identify the top candidates for describing the population as a mathematical function.

Because the subject data represented physical phenomena20, for which negative values

were not physically possible in some attributes21, analyses took care as appropriate to

override the default @Risk settings which allow distributions with both negative and

positive numbers in the output. The @Risk settings were adjusted as necessary to avoid

negative tails where they were physically impossible.

As is common with physical phenomena, the bestfit distribution functions tended to favorLogNormal, LogLogistic or Weibull distributions and less commonly a Normal (Gaussian)

distribution. The Weibull and other logbased distributions are especially applicable to

reliability analyses22. For a given data set, the best fit as indicated by AndersonDarling

criteria was adopted for further analysis although the top three distribution candidates

were used to estimate uncertainties (error bars) in the adopted distribution.

NaturalGasInfrastructureAnalysis

Black & Veatch utilized RBACs GPCMTM model as a basis to analyze the ERCOT and

surrounding regions natural gas market infrastructure. The GPCMTM model operates using

an algorithm to solve for optimal equilibrium price and quantities by balancing multiple

demand and supply nodes in the market. As a network model, GPCMTM

nodes representproduction regions, pipelines, storage facilities, and enduse customer groups. Black &

Veatch supports GPCMTM with a detailed database of proprietary and public sources that

was modified to support the assumptions and scenarios for this study.

The GPCMTM model balances supply and demand from all the regions to find an equilibrium

solution that maximizes producer profit and minimizes consumer cost. Based on Nobel

Prizewinning economist Paul Samuelsons theory, the economically efficient, market

clearing solution will dispatch lower cost supplies before more expensive ones and

customers willing to pay more will be served before those willing to pay less. As shown in

Figure 12, quantity (Q) supplied to market grows as price (P) rises from point of production

19Anderson-Darling is one of several commonly applied tests for measuring the goodness-of-fit of

distribution functions applied to real data. Compared with alternative methods, Anderson-Darling has been

found to provide better performance for distributions with extensive tails.20 Physical phenomena studied here included both integer and decimal numbers. Integer representations

included presence/absence of an incident or event. Decimal numbers included temperature, wind speed,

heating-degree days or system impacts such as Dth/d or MW lost.21 Negative values are possible for temperature but not for other attributes studied, including wind speed,

HDDs, Dth/d, MW and numbers of incidents or events.22 See Meyer S. L. (1975), op cit.


34/107



(Ps) to point of consumption (PD) until the cost of transportation exceeds market price and

supply retracts. Namely, supplies from the supply region will continue to be transported to

the consumption regions until either the price differentials between the two regions drops

below the transportation cost or the transportation capacity between the two regions is

exhausted. The resulting prices, consumption and production quantities represent market

equilibrium.

Demand Rich Region Supply Rich Region

Figure12SupplyDemandFundamentals.

One of the challenges of understanding the risk of gas curtailment to electric generators

within ERCOT is to determine the demand placed on the pipelines serving these electric

generators by other sources residential, commercial, and industrial demand within

ERCOTs region as well as residential, commercial, industrial and electric demand from

outside ERCOTs region that are served by the same pipelines. By representing the entire

natural gas infrastructure within North America, the GPCMTM model offers an efficient and

effective methodology to model the impact of the total demand on the pipeline networkfrom other sources within and outside of ERCOTs region. The fundamental model

represents both interstate and intrastate pipeline segments.

Black & Veatch utilized GPCMTM to assess the constraints on the natural gas infrastructure,

represented as a network within the supply/demand model, in responding to demand from

the electric generation sector within ERCOT under the different defined scenarios. For each

scenario, a corresponding estimate of demand, supply and any applicable scenariospecific

infrastructure constraints were defined.

IntegratedMarketModeling

Black & Veatch has developed an Integrated Market Modeling (IMM) process which is usedto prepare its integrated longterm view on energy markets, the Energy Market Perspective

(EMP). In order to arrive at this market view, Black & Veatch draws on a number of

commercial data sources and supplements them with our own view on several key market

drivers, for example, power plant capital costs, environmental and regulatory policy, fuel

basin exploration and development costs, and gas pipeline expansion.

S

Q

P D

S

Q

P D

Q

P D

Q

P

S

D

Q

P

S

D

S

Q

P D

S

Q

P D

Q

P D

Q

P

S

D

Q

P

S

D

TransportationCosts

PD

PS

S

Q

P D

S

Q

P D

Q

P D

Q

P

S

D

Q

P

S

D

S

Q

P D

S

Q

P D

Q

P D

Q

P

S

D

Q

P

S

D

TransportationCosts

PD

PS

TransportationCosts

PD

PS


35/107



Figure13Black&VeatchIntegratedMarketModelingProcess

EMP is an integrated view of natural gas and power markets across North America, and the

northern portion of Baja California, Mexico, that is electrically interconnected to the U.S.

The study period of 10 years is marked by expectations of significant growth in the use of

natural gas for electric generation in North America driven by environmental policies and

resulting coal retirements and the cost competitiveness of natural gas technology withother fuel sources on a fixed and variable cost basis. By providing a careful consideration of

the multiplicity of factors impacting todays energy markets, the Black & Veatch EMP uses

an integrated market analysis process to arrive at a comprehensive view of how the energy

world can evolve from todays starting point, providing a sound framework for decision

making. The EMP was utilized to provide underlying assumptions for this study.

5.3GLOBALASSUMPTIONS

To evaluate risks of gas curtailments which could impact power generation within ERCOT,

Black & Veatch made assumptions which were necessary to enable objective analyses

within a reasonable scope. Assumptions which apply to all aspects of the study were as

follows:

Natural gas markets in Texas operate efficiently and economically during theanalysis period and market liquidity or mandated curtailments do not comprise

limiting factors contributing to risk of curtailment of gas supply to electric

generators. Appendix E provides a detailed review of the liquidity of the natural gas

market in ERCOT to support this assumption.

Natural gas supply is projected to grow during the analysis period in the Lower48as shown in Figure 14 with growth in unconventional natural gas production led by

shale gas offsetting declines in conventional natural gas production.


36/107



Figure14 Lower48NaturalGasSupplyProjection

Natural gas supply in Texas is projected to flatten with declines in conventionalproduction being offset by growth in unconventional production, primarily from the

Barnett Shale and Eagle Ford shale plays as shown in Figure 15.

Figure15 TexasNaturalGasSupplyProjections

Natural gas demand in the Lower48 is projected to grow over the analysis period asshown in Figure 16. Growth in gas demand for electric generation is the primary

0

10

20

30

40

50

60

70

80

2008 2010 2012 2014 2016 2018 2020

Bcf/d

Lower48NaturalGasSupplyProjections

Conventional Shale CoalBedMethane LNG

0

2

4

6

8

10

12

14

16

18

20

2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Bcf/d

Conventional Barnett EagleFord Haynes ville GraniteWash GOM&StateWaters

TexasNaturalGasSupplyProjections


37/107



driver for growth in natural gas demand as environmental regulations and lower

gas prices lead to an increased share for natural gas in the electric generation mix.

Figure16 Lower48NaturalGasDemandProjection

Natural gas demand in Texas is projected to increase as demand for natural gas forelectric generation increases as shown in Figure 17. Gasfired generation capacity

within ERCOT is projected to increase by over 15,000 MW in the next 10 years and

natural gas share of electric generation within ERCOT is projected to increase to

50% over this time period.

0

10

20

30

40

50

60

70

2008 2010 2012 2014 2016 2018 2020

Bcf/d

Residential Commercial Industrial Electric

Lower48NaturalGasDemandProjections


38/107



Figure17 TexasNaturalGasDemandProjection

Natural gas pipelines and storage facilities are represented in detail in thefundamental supplydemand model used in this study. In order to examine the

sufficiency of the pipeline grid to serve the demand of electric generators within

ERCOTs service region, electric generation facilities were grouped together on the

basis of the natural gas pipelines serving them and their locations and linked (as

demand nodes) to the pipeline network.

Accuracy and precision of statistical analyses are limited by available curtailmentevent data. The different curtailment scenarios considered each offered different

levels of data availability and allowed varying precision in the analytical effort of

this study.

This study takes a conservative view on the risk of gas curtailment to electricgenerators. No mitigating measures have been incorporated in developing the

results presented here. In reality, when natural gas supply or delivery is impacted,

the redundancy and interconnectedness in the natural gas market generally

provides consumers (including electric generators) with alternate sources and

routes for natural gas supply to partially or fully serve their needs. Pipeline

linepack, natural gas storage and displacement of supply from other markets could

all contribute to mitigate the risk of disruption of natural gas supply to electric

generators within the ERCOT service region that are presented in this study.

0

1

2

3

4

5

6

7

8

9

10

2008 2009 2010 2011 20 12 2013 2014 201 5 2016 2017 2018 20 19 2020 2021

Bcf/d

Residential Commercial Industrial Electric

TexasNaturalGasDemandProjections


39/107


BLACK&VEATCH|6.0 RiskAssessment Results 35

6.0 RiskAssessment Results

6.1 IMPLICATIONSFROMFREEZINGWEATHER

6.1.1. ProbabilisticAnalysis

AnalysisMethodology

E

BV ERCOT Gas Study Report March 2012

Documents

Transcript of BV ERCOT Gas Study Report March 2012